Skip to main content

Welcome to OAK

Welcome to the OAK documentation! OAK is an open-source platform designed to make building, customizing, and deploying AI agents effortless. With its modular design and built-in integrations, you can connect to any LLM, extend functionality with powerful plugins, and seamlessly embed AI into your workflows.

OAK empowers developers with flexibility and scalability, allowing you to tailor every aspect of your AI solutions. Dive in and unlock the full potential of OAK to bring your vision to life—without the hassle.

Basic Setup

Select this option to run the app locally without modifying the codebase or developing custom plugins.

  1. Start the App: Execute the following command to launch the app:

    npx @open-agent-kit/cli run docker

  2. Configure the Model: The CLI will prompt you to select a model and provide the associated API key. See the Configuration section below for more information.

  3. Access the App: Open your browser and navigate to:

    http://localhost:3000

Advanced Setup

Select this option if you wish to modify the default model, install plugins, or develop your own plugins.

Scaffold a New Project​

To begin, scaffold a new project by running the following command:

npx @open-agent-kit/cli create project

This command will generate a new project with the default configuration.

Set Up a Database​

You can set up a PostgreSQL database in a Docker container using the command below. Replace the placeholder values with your actual configuration:

docker run --name <container_name> \
  -e POSTGRES_USER=<your_username> \
  -e POSTGRES_PASSWORD=<your_password> \
  -e POSTGRES_DB=<your_database> \
  -p 5432:5432 -d pgvector/pgvector:pg17

Alternatively, for a simpler setup and maintenance, consider using a managed database service like Neon.

Configuration​

Obtain API Keys​

This project utilizes the Vercel AI SDK to initialize and interact with the respective AI models. Ensure that the environment variables are correctly set to configure the AI model. For more information, refer to the Vercel AI SDK Providers documentation.

Generate App Secret Key​

You can generate a secure app secret key using the following tool:

https://api.open-agent-kit.com/generate-secret.html

Set Environment Variables​

Duplicate the .env.example file and rename it to .env. Set the necessary variables in the .env file according to your setup.

cp .env.example .env

Ensure all required variables are correctly configured to match your environment and API keys.

Detailed Environment Variables Description​

  • App Secret Key:

    • APP_SECRET: Secret key for your application. Generate using: openssl rand -hex 32.

  • Base URL:

    • APP_URL: The base URL for your application, default is http://localhost:5173.

  • PostgreSQL Database Connection:

    • POSTGRES_PASSWORD: Password for PostgreSQL database.
    • DATABASE_URL: PostgreSQL database connection string, including user, password, host, port, and database name. The schema=public specifies the schema to use.

  • Google Authentication:

    • GOOGLE_CLIENT_SECRET: Client secret for Google login authentication. Obtain this from the Google Cloud Console by creating OAuth 2.0 credentials.
    • GOOGLE_CLIENT_ID: Client ID for Google login authentication. Obtain this from the Google Cloud Console by creating OAuth 2.0 credentials.

  • AI Service API Keys:

    • OPENAI_API_KEY: API key for accessing OpenAI's services.
    • GOOGLE_GENERATIVE_AI_API_KEY: API key for Google's Generative AI services.
    • ANTHROPIC_API_KEY: API key for Anthropic's services.
    • XAI_API_KEY: API key for XAI's services.

  • Email Service Configuration (SMTP):

    • SMTP_HOST: Hostname of the SMTP server.
    • SMTP_PORT: Port number for SMTP, typically 587 for TLS.
    • SMTP_USER: Username for SMTP authentication.
    • SMTP_PASSWORD: Password for SMTP user account.
    • SENDER_EMAIL: Email address used as the sender in email communications.

Local Development​

Step 1: Install Dependencies​

First, navigate to the project directory. Then, install the necessary dependencies by running:

npm install

Step 2: Run Initial Migration​

Ensure that your database is running and accessible. Apply the initial database migrations using Prisma with the following commands:

npm run generate
npm run migrate:deploy

Step 3: Start the Application​

To start the application, execute:

npm run dev

Once started, the admin interface will be accessible at http://localhost:5173/.